CN1299223C - Providing a consistent hierarchical abstraction of relational data - Google Patents

Abstract

A hierarchy may be explicitly or implicitly reflected in existing data maintained external to hierarchy structures. Such hierarchies are referred to herein as a 'pre-existing hierarchies'. Techniques are described herein for capturing a pre-existing hierarchy in hierarchy structures. After a pre-existing hierarchy has been captured, there exist two independent sets of data that reflect the hierarchy: the external hierarchy definition and the internal hierarchy definition. Changing either hierarchy definition changes the hierarchy. Therefore, to maintain a consistent reflection of the hierarchy, the internal hierarchy definition must be modified in response to changes made to the external hierarchy definition, and the external hierarchy definition must be modified in response to changes made to the internal hierarchy definition. Various techniques are described for maintaining consistency between the two hierarchy definitions.

Description

The consistent hierarchical abstraction of relation data is provided

Related application

The present invention requires the right of priority of following U.S. Provisional Patent Application, and its full content is hereby expressly incorporated by reference:

The U.S. Provisional Patent Application No.60/326 that submitted by Eric Sedlar and Viswanathan Krishnamurthy September 28 calendar year 2001,052, name is called: " being provided with the visit based on file (File Based Access Provided With a DatabaseSystem) of Database Systems ";

The U.S. Provisional Patent Application No.60/378 that on May 7th, 2002 was submitted to by Nipun Agarwal, Ravi Murthy, Eric Sedlar, Sivasankaran Chandrasekar, Fei Ge, Syam Pannala, Neema Jalali and Muralidhar Krishnaprasad, 800, name is called: " to the SQL visit (SQL Access to Data that Provides a File System Abstraction) of data that the file system abstract is provided ".

The invention still further relates to following U.S. Patent application, its full content also is hereby expressly incorporated by reference:

The U.S. Patent application of submitting to by Nipun Agarwal, Ravi Murthy, Eric Sedlar, Sivasankaran Chandrasekar and Fei Ge on the same day with the application, sequence number is: _ _ _ _ _ _ _ _ _ _ _, name is called: " being used for the manipulater (OPERATORS FOR ACCESSINGHIERARCHICAL DATA IN A RELATIONAL SYSTEM) at relational system access individual-layer data " (attorney docket number: No.50277-1975);

The U.S. Patent application of submitting to by Ravi Murthy, Muralidhar Krishnaprasad, Sivasankaran Chandrasekar, Eric Sedlar, Vishu Krishnamurthy and Nipun Agarwal on the same day with the application, sequence number is: _ _ _ _ _ _ _, name is called: " being used for the mechanism (MECHANISM FOR MAPPING XML SCHEMAS TO OBJECT-RELATIONAL DATABASE SYSTEMS) of XML mode map to object-relational database system " (attorney docket number: 50277-1977);

The U.S. Patent application of submitting to by Nipun Agarwal, Eric Sedlar and Ravi Murthy on the same day with the application, sequence number is: _ _ _ _ _ _ _ _ _, name is called " index (INDEXING TOEFFICIENTLY MANAGE VERSIONED DATA IN A DATABASESYSTEM) that is used for the modification data of high-efficiency management Database Systems " (attorney docket number is 50277-1978);

The U.S. Patent application of submitting to by Ravi Murthy, Eric Sedlar, Nipun Agarwal and Neema Jalali on the same day with the application, sequence number is: _ _ _ _ _ _ _ _, name is called: " being used to store the content of laminated tissue's resource and the mechanism of attribute (MECHANISMS FOR STORING CONTENT AND PROPERTIESOF HIERACHICALLY ORGANIZED RESOURCES) " (attorney docket number: 50277-1979);

The U.S. Patent application of submitting to by Ravi Murthy, Eric Sedlar, Nipun Agarwal, SamIdicula and Nicolas Montoya on the same day with the application, sequence number is: _ _ _ _ _ _ _ _, name is called: " being used for the mechanism (MECHANISM UNIFORM ACCESS CONTROL IN ADATABASE SYSTEM) in the unified access control of Database Systems " (attorney docket number: 50277-1980);

The U.S. Patent application of submitting to by Syam Pannala, Eric Sedlar, Bhushan Khaladkar, Ravi Murthy, Sivasankaran Chandrasekar and Nipun Agarwal on the same day with the application, sequence number is: _ _ _ _ _ _ _ _, name is called: " but loading attachment (LOADABLE UNITS FORLAZY MANIFESTATION OF XML DOCUMENTS) that the inertia that is used for XML document shows " (attorney docket number: 50277-1981);

With the application on the same day by the U.S. Patent application of Neema Jalali, Eric Sedlar, Nipun Agarwal and Ravi Murthy application, sequence number is: _ _ _ _ _ _ _ _, name is called: " index is used to provide the mechanism (MECHANISM TO EFFICIENTLY INDEXSTRUCTURED DATA THAT PROVIDES HIERARCHICAL ACCESSIN A RELATIONAL DATABASE SYSTEM) of the structural data of layering access expeditiously in relational database system " (attorney docket number: 50277-1982).

Technical field

The present invention relates in relational database to store the information of organizing according to certain level, relate in particular to the technology that management is used to catch the metadata of this level.

Background technology

People tend to the category organizational information.The classification of organizational information is the classification of relative to each other organizing with the level of certain form usually itself therein.For example, individual animal belongs to kind, plants to belong to genus, belongs to section, and section belongs to order, and order belongs to guiding principle.

Along with the appearance of computer system, the technology of stored electrons information is developed, has reflected the demand of people to laminated tissue to a great extent.For example, adopt principle realization traditional computer file system usually based on laminated tissue.Particularly, common file system has by the catalogue of level setting and is stored in document (document) in the catalogue.Be that the hierarchical relationship between the catalogue has reflected certain intuitive relationship between the implication of giving catalogue ideally.Equally, be ideally, each document based in the content of document and certain intuitive relationship that is used for storing between the implication that the catalogue of this document is endowed be stored in the catalogue.

Fig. 1 shows the example of typical file system.The file system that illustrates comprises a plurality of catalogues of arranging by level.In catalogue, store two documents 118 and 122.Especially, document 118 and 122 title all are " Example.doc ", are stored in respectively in catalogue 116 and 124, and the title of these two catalogues is respectively " Word " and " App4 ".

In TOC level, catalogue 116 is sub-directories that name is called the catalogue 114 of " Windows ", and catalogue 114 is sub-directories of catalogue 110.Similarly, catalogue 124 is sub-directories that name is called the catalogue 126 of " VMS ", and catalogue 126 is sub-directories of catalogue 110.Catalogue 110 is referred to as " root " catalogue, and reason is that this catalogue is the catalogue of every other catalogue origin.In a lot of systems, symbol "/" is used to refer to root directory.

When pressing the hierarchical organization electronic information, according to each data item that can search information by hierarchy to " path " of the entity that comprises this data item.In a hierarchical file system, begin by root directory to the path of data item, carry out the final catalogue that comprises interested data item that arrives downwards along hierarchical structure of directory.For example, the path to document 118 comprises catalogue 110,114 and 116 successively.

The layering storage system allows different data item to have identical title usually.For example, in file system shown in Figure 1, document 118 and 122 title all are " Example.doc ".Thereby,, except document title, also have other requirement in order to discern given document clearly.

A kind of short-cut method that is used for discerning and locatees the specific data item that is stored in the information in the layering storage system is by using " pathname ".The pathname method is a kind of basis is come unique recognition data item to the path of data item by hierarchy a short-cut method.Pathname comprises a sequence title.Under the situation of file system, each title in this title sequence all is " filename ".Term " filename " refers to the title of directory name and document, and reason is that catalogue and document all are taken as " file ".

In file system, the filename sequence in the given path name starts from the title of root directory, comprises along the title of all catalogues in the path from root directory to interested data item, ends at the title of interested data item.Usually, the directory listing that travel through use certain list separator (for example, "/", " " or "; ") couple together, to produce pathname.Therefore, the pathname of document 118 is :/Windows/Word/Example.doc, and the pathname of document 122 is :/VMS/App4/Example.doc.

Relation marked change between catalogue (file) and the content that comprised thereof along with the difference of the type of system of laminated tissue.In a kind of pattern that numerous embodiments (for example Windows and DOS file system) is adopted, need each file just to have a father, form tree structure.In more complicated pattern (for example in using hardwired unix filesystem), hierarchy adopts the form of oriented graph, and file can have a plurality of fathers.

Organize the top and bottom process of electronic information opposite with being used for, relational database with information storage in the table of forming by row and column.Discern every row by unique RowID.Every row have been represented the attribute of record, and every row has been represented specific record.Submit inquiry to by data base management system (DBMS) (DBMS), from database, obtain data to management database.Inquiry must be followed the database language that data base management system (DBMS) is supported.Structured Query Language (SQL) (SQL) is the example of the database language of a lot of available data base management systems supports.

Every kind of stocking system all has merits and demerits.The stocking system of laminated tissue is simple, directly perceived, be easy to realize, is the mode standard that most application programs are used.Unfortunately, simply laminated tissue does not provide the complex data search operaqtion needed support.For example, must check the content of each catalogue, thereby retrieve all documents of establishment in certain day with particular file name.Because must search for All Files, laminated tissue is not easy to retrieving.

Relational database system is very suitable for storing a large amount of information and with unusual flexible way access data.For the system of laminated tissue, also can from RDS, retrieve easily and efficiently even meet the data of complicated search condition.Yet, customization and to submit to the process of inquiry but only not travel through gradation directory to database server directly perceived, and exceeded the skill level that a lot of computer user grasped.

Past, the system of laminated tissue with concern that organization system realizes in different mutually incompatible modes.Yet,, concern that organization system can simulate the system of laminated tissue by some additional programs.When storage capacity that needs relational system and dirigibility, when needing simultaneously the intuitive of hierarchical system and ubiquity again, especially need this simulation.

Description of drawings

By example explanation the present invention, be not to be used to limit the present invention, in the accompanying drawings, the identical identical parts of label representative, wherein:

Fig. 1 illustrates the block diagram of hierarchical file system;

Fig. 2 illustrates the block diagram of the node level that is associated with the source;

Fig. 3 is the block diagram that can be used to catch the relation table of the level shown in Fig. 2 according to an embodiment of the invention;

Fig. 4 illustrates the block diagram of hierarchy and content structure according to an embodiment of the invention;

Fig. 5 is the block diagram of relation table, wherein by being stored in the non-leaf node that level of information is shown corresponding to the value in the row of leaf node;

Fig. 6 is the block diagram that is stored in the implicit level of data in the table of Fig. 5; And

Fig. 7 is the block diagram that can implement the system of embodiments of the invention.

Embodiment

The invention describes the method and system that is used for storing the source that belongs to level of information in the structure of relational database system.In the following description, for purpose of explanation, thorough of the present invention many details have been illustrated in order to provide.Yet obviously, even without these specific detail, the present invention also can implement.In other example,, known structure and device are shown with block scheme for fear of to unnecessary the bluring of the present invention.

Functional overview

The invention provides the multiple technology of being convenient to managing hierarchically data in relational database system.According to an embodiment, with the storage organization (" content structure (content structure) ") that is used for storing the source that belongs to level be used for storing the storage organization (" hierarchy (hierarchy structure) ") of seizure and separate about the data of level.

In some cases, reflected level in the available data of outside hierarchy, keeping clearly and impliedly.Such level is known as " having levels earlier ".In these cases, need in hierarchy, catch and have levels earlier, so that hierarchy can be used as the optional access path in source.For example, according to routing information, available hierarchy access originator.And, even when by directly sending inquiry when visiting the source to content structure, accessing operation also can obtain the benefit of all features, and the visit structure relevant with hierarchy, for example, any security mechanism and any index are built in the security mechanism in the hierarchy, and index is established to strengthen the operating performance in the source that visits according to the position of source in level.

Each has levels earlier all and following data association, and (1) is present in the data outside the hierarchy, and (2) have reflected first stratified data.Here such data are called " outside level definition ".To have levels earlier in order in hierarchy, catching, must data (referring to " inner level definition " herein) to be added in the hierarchy according to outside level definition.

Caught have levels earlier after, also have the independent datas of two groups of reflection levels: definition of outside level and inner level define.Change any definition and all changed level.Thus, in order to keep the consistance reflection of level,, must change inner level definition in response to the change that outside level definition is made.Described herein and be used to keep two kinds of conforming multinomial technology between the level definition.

In relational database system, store individual-layer data

Fig. 2 illustrates the block diagram of the level 200 that uses in the example, and the example that herein provides is for the ease of understanding the embodiment of the invention.Level 200 comprises eight nodes.The node of highest ranking is known as " root " node in this level.The node of the end of each branch is " leaf " node.Node between root node and the leaf node is " centre " node.In schematic level, node 1,2 and 3 is intermediate nodes, and node 4,5,6 and 7 is leaf nodes.

In level of information, node is corresponding to information.Usually and the item of information that is associated of each node the title of certain form and certain type content will be arranged.For example, in the level corresponding to hierarchical file system, node is usually corresponding to file (here, " file " or " catalogue " is a type of file).Each such file will have the content of title and certain form.

Under many circumstances, the title relevant with node needs not to be unique in level.In the example shown in Fig. 2, node 1 and node 7 all have title " a ".Although their title is identical, they are other different nodes of branch, occupy different positions in level, are associated with diverse content.

A kind of specification is suitable for all (ONE-SIZE-FITS-ALL)

Fig. 3 is the block diagram that can be used to two tables (source table 302 and chained list 350) of the level 200 in the representation relation Database Systems.Source table 302 comprises the row of each node that is used for level.The row that is used for given node comprises title related with given node and and the related data of given node.For example, row 304 is corresponding to node 1, comprises title " a " and the data 306 related with node 1.Source table 302 is examples that a kind of specification is suitable for whole methods, and wherein, the data that are used for all data sources in level are stored in identical structure, and do not consider the data type in these sources.In the example of source table 302, being used for the data structure of storage source data content is single-row 308.The data that are used for these row can be LOB (Large Object, big object) types for example, such as scale-of-two LOB (BLOB) or character LOB (CLOB).

Chained list 350 comprises the delegation of the every pair of father-subrelation that is used for level 200.Father's row 378 are held the father's who represents father-subrelation value, and sub-row 380 are held the value of the son of representing father-subrelation.For example, row 352 expression root nodes are fathers of node 1.Equally, row 354 and row 356 represent that respectively node 1 is the father of node 2 and node 3.

Hierarchy and content structure are separated

Table 302 and 350 all information with relational format seizure level 200.But, if using a kind of specification by the representative of source table 302 to be suitable for whole methods, the content of the source data related with level 200 stores, then can not utilize the ability of relational database system fully.Especially, source table 302 comprises and is used for the resource content related with node is stored in the single-row of level 200.When the source data related with node was simple data type (for example, character string or integer), such saving format was enough usefulness.But under many circumstances, the source data related with node is complicated data type in level, for example, and complicated record or XML document.Under these circumstances, source data stores the user carries out complex search to content the ability that restricted in single-row.Even when these Database Systems provided support to the complicated source data component of search in single-row, such search performance also can suffer damage.

According to one embodiment of present invention, by source contents that will be associated with node and the table separate storage of catching hierarchical information, overcome and stored and search belongs to the relevant problem of complex source data of level of information.

For example, Fig. 4 shows and the similar source of source table 302 table 402.Yet the difference of source table 402 and source table 302 is that source table 402 no longer stores the content of this source data in the row of each source data.Certainly, the DATA in the table 302 of source is listed as by row of the DATA TYPE in the table 402 of source and DATA REF and is listed as this two row replacement.

For the every row in the table 402 of source, DATA TYPE row store indication and the data of the Source Type that the node of the middle identification of being expert at is associated.For example, row 404 is corresponding to the node 1 of level 200.The type that the DATA TYPE of row 404 row illustrate the source related with node 1 is T1.Equally, row 406 is corresponding to node 3, and the DATA TYPE of row 406 row are indicated and the type in the source that node 3 is associated is T5.

For every row of source table 402, DATA REF row are storing and the reference information in the source that the middle node of discerning of being expert at is associated.As use herein, term " reference " is often referred to any data that are used for locating other data in data system.For example, DATA REF row can store pointer or more complicated information, such as " ref " data type of oracle database system support.The present invention is not limited to be used for to locate the data of any particular type in the source that the node with level is associated.

With hierarchy separate storage source contents a lot of benefits are arranged, hierarchy is used for catching and reflects hierarchical relationship between the source data.For example, because source data and source table 402 separate storage, so can wherein store the content structure of source data according to the attribute definition of source data.For example, be the employee record with node 1 related source data, this source data is stored in the table 450 with the row that are used for each employee record's field.On the other hand, are corporate records with node 3 related source datas, this source data is stored in the table 456 with the row that are used for each corporate records's field.

In addition, because hierarchy (for example, table 402 and 350) and content structure (for example, table 450 and 456) separately, so can set up layered system for existing relation data, very little to the influence of data capacity, and do not need to change the existing inquiry that is used to visit these data.For example, employee's table 450 can determine that format is advanced level 200 just to be existed very early.Equally, before decision, can a lot of inquiries and view have been defined for access list 450.Because can create table 350 and 402 and do not influence the existing structure of table 450, can set up and be used for the hierarchy that containing of level 200 is stored in the resource of table 450, and not need access list 450 or be used for existing inquiry of change table 450 and view.

According to an embodiment, can be suitable for whole methods with a kind of specification that use is used for some resources and realize resource table 402, resource table 402 also comprises the reference to other resource contents.For example, though can be stored in the table 450 with node 1 related resource content,, can continue to be stored in the single LOB row of table 402 with node 2 related resource contents.The data type of resource, shown in being listed as at DATA TYPE, can be by the factor of database server as storage resources content in the table of the LOB row that determine whether at resource table 402 or one or more separation.

The level of preexist

Table 450 just can be isolated to hierarchy at first fully and exist, such as resource table 402 and chained list 350.Therefore, hierarchy does not have to catch any information of any data structure that can belong to about the data in the table 450 at first.Yet the information that is stored in such hierarchy outside can be set up level clearly or impliedly between each row of table 450.

Under the situation of table 450, impliedly set up such level by the information that is included in EMD_ID and the MANAGER row.Especially, the data in the EMD_ID of table 450 and the MANAGER row are set up implicit " management hierarchy ", and wherein: the employee related with emp_id1 is directly under the employee related with emp_id 3; The employee related with emp_id 2 is directly under the employee related with emp_id 1; And the employee related with emp_id 3 without any the supvr (from but the root node of management hierarchy).

In present example, management hierarchy be the level of preexist and be stored in EMD_ID and MANAGER row in data comprise exterior layer time definition.According to management hierarchy, use the data in the hierarchy access list 450, must in layered sheet, catch information about management hierarchy.After this multiple technologies that are used to catch about the information of the data structure of preexist will be described in detail.

Catch the level of preexist

According to an embodiment, the hierarchical information of catching preexist relates to: (1) stores each node that row are used for the level of preexist in the table of source, and the information of (2) father-subrelation between the node of the level of the middle storage seizure of hierarchy (for example, hierarchical index) preexist.In addition,, so, metadata must be increased in the hierarchy, be established as the son of the node in the larger data structure with root node the level of preexist if the level of preexist is advanced by grafting in the bigger level that has reflected in hierarchy.

The operation of carrying out following task is quite direct: (1) stores the row of each node of the level that is used for preexist in resource table, and the metadata of node that (2) will be used for the level of grafting preexist increases to bigger level.On the other hand, according to the attribute of outside level definition, can change the technology of the father-subrelation between the node of the level that is used to catch preexist.For example, if SQL contextual definition father-subrelation so, adopts and uses suitable connect-by clause's sql command can catch father-sub-information.As selection, if the field relevant according to level, father-subrelation implies, and so, adopts enumeration method can catch father-sub-information.Will be discussed in more detail below connect-by and enumeration method.

The automatic seizure of the level of preexist

According to an embodiment, database is configured, and with in response to certain information that receives about the level of preexist, automatically performs preexist level capture operation.Particularly, according to an embodiment, database is configured, to be used for discerning and the information of the content structure of the level association of preexist and (2) are used in reference to and are shown in the connection standard of setting up the standard of father-subrelation between the resource that the certain content structure comprises, automatically carry out the level capture operation of preexist in response to receiving (1).

For purpose of explanation, suppose the management hierarchy of wishing that seizure is related with table 450.In this case, are EMD_ID is equal to the row of the MANAGER that stipulates in given row sons in response to data of reception (1) Identification Lists 450 and the given row of (2) indicating gauge 450, can carry out automatic capture operation.Can will be described below according to the technology that this information is used for catching automatically the level of preexist.

Use CONNET BY to catch father-subrelation

As above-mentioned, the father-subrelation of the level of some preexists is illustrated by the SQL relation.For example, set up the management hierarchy of the record in the table 450 by the data in the MANAGER row that are included in table 450.In this case, the father-subrelation of the level of preexist can adopt appropriate C ONNECT BY clause's database command to catch by using.

For example, the automatic capture operation of management hierarchy is carried out by following action: (1) creates row in resource table 402, with the node of representative corresponding to the row in the table 450; And (2) comprise that by execution CONNECT BY clause's SQL statement produces the metadata about the hierarchical relational between those nodes.Every provisional capital of node that representative is used for the particular row of table 450 in the resource table 402 has the reference to this particular row.For example, row 404 representatives of resource table 402 are used for the node of the row 420 of table 450, thereby comprise the reference to row 420.

About the routing information that generates and the level of preexist is associated, suppose that table 450 is named as ' EMP ' and is defined to comprise following row: (emp_id integer, first_namevarchar2 (80), last_name varchar2 (80), age integer, manager integer).In these cases, use CONNECT BY relation (for example, ' START WITHmanager IS NULL CONNECT BY PRIOR manager=emp_id ') to generate clauses and subclauses trees (according to data, variable depth).In this case, the user also can specify train value with the title (i.e. ' first_name ') as path elements usually.

Use enumeration method to catch the level of preexist

The level of field implicit difiinition preexist that in some cases, can be by having hierarchical relationship mutually.For example, consideration comprises the relation table of the product tabulation of an offering for sale.For example, such relation table can be as giving a definition:

CREATE?TABLE?PRODUCT_INFO

(

manufacturer?VARCHAR2(128)

product_type?VARCHAR2(48)

product_name?VARCHAR2(80)

sku?NUMBER

)；

Such table can be filled, and is for example, shown as the table 500 among Fig. 5.In this example, level is implied in the hierarchical relational between MANUFACTURER, PRODUCT_TYPE, PRODUCT_NAME and the SKU field.Especially, each SKU value is in a lot of potential SKU value related with specific PRODUCT_NAME value one.Each PRODUCT_NAME value is in a lot of potential PRODUCT_NAME value related with specific PRODUCT_TYPE value.Each PRODUCT_TYPE value is in a lot of potential PRODUCT_TYPE value related with specific MANUFACTURER value.

" products catalogue level " implicit in the value that table 500 comprises is by Fig. 6 explanation.Referring to Fig. 6, each in four layering relevant fields of table 500 is all corresponding to stratum level, and wherein, the MANUFACTURER field is highest ranking (only under a root node), and the SKU field is the lowest class (representative " leaf " node).

When defining the level of preexist by this way, can use enumeration method to catch father-subrelation.Especially, row or the multiple row that comprises the enumerator that is involved in the level is used to generate " resource set node (collection node) ".Each value of every row all is new " resource set " in a certain degree of depth of level, and every row all are used to the new grade of level.The row that provide at last are taken as file, but not resource set.

Especially, by the value that is included in the MANUFACTURER row is carried out the DISTINCT operation, can set up the node in the first estate of level.Be used for MANUFACTURER and PRODUCT_TYPE values of all row by connection, and the value of the generation after connecting carried out DISTINCT operate, can set up the node in second grade of level.Equally, be used for MANUFACTURER, PRODUCT_TYPE and PRODUCT_NAME values of all row by connection, and the value of the generation after connecting carried out DISTINCT operate, can set up the node in the tertiary gradient of level.At last, be used for MANUFACTURER, PRODUCT_TYPE, PRODUCT_NAME and the SKU value of all row by connection, and the value of the generation after connecting carried out DISTINCT operate, can set up the leaf grade of level.

The title of supposing the node of given grade is not unique, carries out above-mentioned attended operation.For example, the product of supposing PRODUCT_TYPE value PD1 can have the identical title of product with PRODUCT_TYPE value PD2.Yet,, so, can obtain node by the value of PRODUCT_NAME row is carried out the DISTINCT operation simply in the PRODUCT_NAME of level level if the title of PRODUCT_NAME row is required it is unique.

Should be noted that by being stored in the capable value of relation table of holding the resource that is used for node and represent the associated pathway title of each leaf node.For example, row 502 (Fig. 5) holds the resource that is used for node 602 (Fig. 6) and corresponding to the value of the row of the row 502 of associated pathway/MF1/PD2/PN3/SKU3.Therefore, when the period of the day from 11 p.m. to 1 a.m of seeking specific node (for example MF1/FD2), the row that are used for this node impliedly are used for chooser (therefore, the SQL that is used for seeking the son of MF1/PD2 will be ' SELECT FROM PRODUCT_INFO WHEREmanufacture=' MF1 ' and ' product_type '=' PD2 ").

Keep consistance

As mentioned above, after formerly the level of Cun Zaiing has been captured, there are two groups of independent datas of this level of reflection: outside level definition and the definition of inner level.In order to keep the consistance reflection of level,, must revise inner level definition in response to the change that outside level is made.

For example, in case catch, and the management hierarchy of table 450 association is reflected in the relation between MANAGER that (1) be stored in data in the EMP_ID row of the data of hierarchy and (2) table 450 and table 450 data in being listed as.Therefore, can finish in the following manner the change of management hierarchy: (1) changes the EMP_ID of table 450 or the value in the value in the MANAGER row or (2) change hierarchy.In order to keep the consistance of the data in data in the hierarchy and EMP_ID or the MANAGER row, need create a mechanism, make change to one group of data can cause corresponding modify automatically to another group data.

According to an embodiment, change table 450, to make, also can make these inspections and be used for traversal list 450 as some inspection (for example, security inspection) to the part of the tour of resource table 402.(Data ManipulationLanguage, DML) trigger is so that insertion, renewal or deletion action that his-and-hers watches 450 carry out reflect in the data of hierarchy to create data manipulation language (DML) on table 450.For example, the row in the delete list 450 is if desired then checked and is guaranteed deleted capable can not cause any " unsettled row (danglingrows) ".For example, in management hierarchy, are fathers of the node related with row 422 with row 420 related nodes.Therefore, deletion row 420 will make row 422 and management hierarchy separately, thereby cause row 422 " unsettled ".Because unsettled row no longer is the part of level, deletion row 422 will cause from hierarchy the relevant data of deletion and row 420 and row 422.

Equally, in the time will going insertion table 450, thereby the new route that the information of hierarchy will be updated to creating increases clauses and subclauses.Note, when inserting newline, then may allow to visit the subtree that to visit before this.For example, after supposing to delete row 420, insert row 420 once more.By inserting row 420, row 422 quilts are grafting return pipe reason level impliedly.Therefore, the insertion of row 420 causes the data in the hierarchy to be updated, to comprise node and the path that is used for row 420 and row 422.

For keeping the consistance between inner level definition and the definition of outside level, need to carry out accurate operation, these accurate operations change with multiple factor.The influence that must make what action is to keep the attribute that conforming a kind of factor is outside level definition.For example, concern his-and-hers watches 450 application management levels according to SQL, and according to enumerating his-and-hers watches 500 application management levels.When deletion related with the leaf node in the management hierarchy when capable, common, corresponding to the row in the resource table of this node also with deleted.Yet if the deletion the row related with leaf node of products catalogue level, (1) deletes in the resource table row corresponding to this node, and (2) determine whether the father of deleted leaf node has any remaining son.If the father of deleted leaf node is without any remaining son.Then delete in the resource table row corresponding to this node.If deletion, determines then whether the father of this father node has any remaining son corresponding to the row of father node.If the father of this father node without any remaining son, then deletes the row that is associated with father father node resource table equally.Repeat this process, delete all s.m.p nonleaf nodes.

Equally, for relate to abandoning and truncated operation of all row in the delete list, carry out be similar to above-mentioned be used for deletion action be used to keep conforming operation.

In above-mentioned example, the DML trigger is according to the insertion of carrying out on content structure, renewal and the deletion action execution change to the data in the hierarchy.Can realize for example DML trigger of " INSTEAD OF " trigger.Yet,, must create trigger equally and cause the data in the content structure are operated accordingly to guarantee insertion, renewal and deletion action that the data in the level are carried out in order to guarantee consistance.

Ardware overview

Fig. 7 is one and has described the block diagram of realizing the computer system 700 of specific embodiments of the invention.Computer system 700 comprises that bus 702 or other are used to the communicator of the information that transmits, and is connected the processor 704 that is used for process information with bus 702.Computer system 700 also comprises a primary memory 706, and for example random-access memory (ram) or other dynamic memory are connected with bus 702, the instruction that is used for store information and will be carried out by processor 704.The intermediate variable that primary memory 706 also can be used to store temporary variable or produce in the process of processor 704 execution commands.Computer system 700 also comprises ROM (read-only memory) (ROM) 708 or other static storage device that links to each other with bus 702, the instruction that is used to store static information He is used for processor 704.Memory device 710 is provided, and for example disk or CD are connected to bus 702, are used for store information and instruction.

Computer system 700 can be connected to display 712 (for example cathode ray tube (CRT)) by bus 702, is used for display message and gives the computer user.One comprises the input media 714 of letter and number and other key, is connected to bus 702, is used for sending information and Instruction Selection to processor 704.The user input apparatus of another kind of type is a cursor control 716, and for example mouse, trace ball or cursor direction key are used for direction of transfer information and command selection and give processor 704, and is used to control cursor moving on display 712.Input media has two degree of freedom of two axles usually, and one is first (for example x), and one is second (for example y), and this just allows this device to determine the position in the plane.

The present invention relates to the application of computer system 700, be used to implement technology described herein.According to one embodiment of present invention, be included in the processor 704 of one or more instructions of the one or more sequences in the primary memory 706 in response to execution, can realize these technology by computer system 700.These instructions can be read the primary memory 706 from another computer-readable medium (for example memory storage 710).Execution is included in the instruction sequence in the primary memory 706, makes processor 704 carry out treatment step described herein.In optional embodiment, hardware connecting circuit can be used to replace or combine with software instruction and realize this invention.Therefore, embodiments of the invention are not limited in the combining of any specific hardware circuit and software.

The term of here using " computer-readable medium " refers to any medium that participation provides instruction to be used to carry out to processor 704.Such medium has various ways, includes but not limited to: non-volatile media, Volatile media, transmission medium etc.Non-volatile media comprises CD or disk for instance, as memory device 710.Volatile media comprises dynamic storage for instance, and for example primary memory 706.Transmission medium comprises concentric cable, copper cash or optical fiber, also comprises the wiring that constitutes bus 702.Transmission medium also can adopt the form of sound wave or light wave, sound wave that for example produces in the radiowave and the light wave in the infrared data communication.

The common form of computer-readable medium comprises for instance: floppy disk, flexible plastic disc, hard disk, tape, or other any magnetic medium, CD-ROM, other any optical medium, card punch, paper tape, other any physical medium with sectional hole patterns, ROM (read-only memory) (RAM), programmable read-only memory (prom), and Erarable Programmable Read only Memory (EPROM), flash memory (FLASH-EPROM), any other memory chip or cartridge, carrier signal hereinafter described, perhaps other any computing machine medium that can read.

Various forms of computer-readable mediums can be used to one or more instructions with one or more sequences and send processor 704 to and be used for carrying out.For example, instruction at first is loaded on the disk of far-end computer.This far-end computer may send instruction by modulator-demodular unit in phone line these instruction load in its dynamic storage.The modulator-demodular unit of computer system 700 this locality can receive data by telephone wire, and uses infrared transmitter that data-switching is become infrared signal.Infrared eye can receive the data of carrying in the infrared signal, and suitable circuit can pass to bus 702 to data and get on.In primary memory 706, processor 704 is fetched data and execution command to bus 702 again from primary memory data load.The instruction that primary memory 706 receives is stored on the memory device 710 after can or carrying out before processor 704 is carried out selectively.

Computer system 700 also comprises a communication interface 718 that links to each other with bus 702.Communication interface 718 provides a kind of and has been connected to the bidirectional data communication of the network link 720 of LAN 722.For example, communication interface 718 can be Integrated Service Digital Network(ISDN) (ISDN) card, or modulator-demodular unit, provides data communication to connect to the telephone wire of respective type.As another example, communication interface 718 can be the Local Area Network card, and the data communication that is provided to compatible LAN (Local Area Network) connects.Also can realize wireless connections.In any implementation, communication interface 718 can both send and receive electric signal, and electromagnetic signal or light signal, these signaling bearers the digital data stream of the various information of representative.

Network link 720 generally provides data communication by one or more networks to other data equipment.For example, network link 720 can provide by LAN (Local Area Network) 722 to main frame 724 or by the connection of the data set of ISP (ISP) 726 operation.ISP 726 again by ten thousand dimension packet data communication network, is commonly referred to " internet " 728 conversely now, and data communication services is provided.LAN (Local Area Network) 722 and internet 728 all use electric signal, electromagnetic signal or the optical signalling of carrying digital data stream.Through the signal of diverse network, the signal on the network link 720 and through the signal of communication interface 718, carrying and delivering to and, be the demonstration form of the carrier wave of transmission information from the numerical data of computer system 700.

Computer system 700 can be passed through network, network link 720 and communication interface 718 and send message and receive data, comprises program code.In the Internet example, server 730 can pass through internet 728, ISP 726, LAN (Local Area Network) 722 and communication interface 718 and transmit requested application code.

After receiving code, the code that receives can be carried out by processor 704, or is stored in memory storage 710 or other nonvolatile memories and is used for carrying out subsequently.By this way, computer system 700 can obtain the application code of carrier format.

In above-mentioned instructions, the present invention has been described with reference to specific embodiment.Yet, obviously can make various variations and change, and not deviate from the spirit and scope of broadness of the present invention it.Therefore, instructions and accompanying drawing only are used for illustrating and non-limiting the present invention.

Claims (14)

1. method that is used for the data of management database system may further comprise the steps:

Receive at the database server place:

(a) data, described data are used for one or more relation tables are identified as the table that is used for storing the data that are associated with resource content, and described resource content belongs to level of information; And

(b) connection standard, described connection standard is used for indicating the relation of what configuration information level;

Wherein, described one or more relation table comprises at least one table, certain data that described table is used to store the data that are associated with described resource content and plays the effect of outside level definition, described outside level definition is used for setting up hierarchical relationship between described resource;

Use is used to identify the described data and the described connection standard of described one or more relation tables, by carrying out following steps, catch the hierarchical relationship in the described level: carry out one or more inquiries, to retrieve described certain data with respect to described one or more relation tables; And

According to described certain data and described connection standard, in one or more hierarchical structures of from described one or more relation tables, separating, store the information that is used to catch described level of information as inner level definition; And

Use described one or more hierarchical structure and routing information to visit described resource.

2. method according to claim 1, wherein, the step that described storage is used for catching the information of described level of information is included in the row that resource table stores each node that is used for described level of information.

3. method according to claim 1, wherein, the step that described storage is used for catching the information of described level of information is included in the information that one or more hierarchical structures store the father-subrelation between the described resource of expression.

4. method according to claim 1, wherein

Resource in the described level of information is corresponding to the row in the relation table; And

With the father-subrelation between the described resource of value representation in two or more row that are stored in described relation table.

5. method according to claim 4, wherein, the step of the one or more inquiries of described execution comprises carries out the inquiry that is used to specify the CONNECT BY relation between described two or more row.

6. method according to claim 5, wherein:

Described two or more row comprise first row that contain the data that are useful on the son in the expression relation and the secondary series that contains the data that are useful on the father in the described relation of expression; And

The step of described execution inquiry comprises carries out the inquiry that comprises following expression form: START WITH x IS NULL CONNECT BY PRIOR x=y;

Wherein, x represents described secondary series, and y represents described first row.

7. method according to claim 1, wherein:

The resource related with the leaf node in the described level of information is corresponding to the row of relation table; And

Will be in described level of information be arranged in given leaf node on the value that is associated of nonleaf node be stored in row corresponding to described given leaf node.

8. method according to claim 7, wherein:

The described value related with nonleaf node is stored in one or more row of described relation table; And

The step that described storage is used for catching the information of described level of information comprises: will be stored into described one or more row from the row that is used for each different value corresponding to the row of the leaf node of described level of information at resource table.

9. method according to claim 1, further comprising the steps of: as to set up trigger,, upgrade the data that constitute described inner level definition with in response to the change that the data that constitute described outside level definition are made.

10. method according to claim 1, further comprising the steps of: as to set up trigger,, upgrade the data that constitute described outside level definition with in response to the change that the data that constitute described inner level definition are made.

11. method according to claim 1 also comprises the conforming step of keeping automatically between described outside level definition and the described inner level definition.

12. method according to claim 11 is wherein saidly kept conforming step automatically and be may further comprise the steps:

Detection to the operation of described one or more tables when delete with described level of information in the related resource of node;

In response to detecting described operation, automatically perform following steps:

Deletion is corresponding to the row of described node in resource table;

Determine whether remove described node from described level of information causes one or more other nodes unsettled; And

Cause one or more other nodes unsettled if remove described node from described level of information, so, deletion is corresponding to the row of described one or more other nodes in described resource table.

13. method according to claim 11 is wherein saidly kept conforming step automatically and be may further comprise the steps:

Detection to the operation of described one or more tables when delete with described level of information in the related resource of leaf node;

In response to detecting described operation, automatically perform following steps:

Deletion is corresponding to the row of described node in resource table;

Determine from described level of information, to remove the father node whether described leaf node make described leaf node and do not have remaining child node; And

Do not have remaining child node if remove the father node that described leaf node makes described leaf node from described level of information, so, deletion is corresponding to the row of described father node in described resource table.

14. method according to claim 11 is wherein saidly kept conforming step automatically and be may further comprise the steps:

Detection to the operation of described one or more tables when insert with described level of information in the related resource of node;

In response to detecting described operation, automatically perform following steps:

In resource table, insert row corresponding to described node;

Determine whether to cause one or more child nodes of described node are increased to described level of information from the increase of the described node of described level of information; And

If the increase from the described node of described level of information causes one or more child nodes of described node are increased to described level of information, in described resource table, insert row so corresponding to described one or more child nodes.

Images